迭代重建技术对低辐射剂量上肢血管内瘘CT造影成像质量的影响

Application of iterative reconstruction in upper limb intravascular fistula CT angiography

目的 探讨第二代双源CT上肢血管内瘘造影成像中应用迭代重建时，与常规剂量滤波反投影（Filtered back projection，FBP）重建图像质量相当的最适合的低剂量水平。方法 本试验为前瞻性研究，将150例拟行上肢内瘘血管CT造影成像（CTA）检查的患者按照同期随机原则分为常规对照组和4个实验组，每组30例。常规对照组采用常规剂量FBP重建，4个低辐射剂量的实验组应用正弦波图形法迭代重建，管电流分别逐级降低至常规对照组管电流（110 mAs）的70%、60%、50%和40%。取锁骨下动脉近肩关节层面以及桡动脉与头静脉吻合口相近层面作为两个感兴趣区，采用方差分析比较各组图像质量的客观评价指标[信噪比（SNR）、对比噪声比（CNR）]以及辐射剂量（容积CT剂量指数、剂量长度乘积、有效剂量）的差异。图像质量（各组的管腔对比度、边缘锐利度、主观噪声及图像总体质量）的主观评价采用4分法，采用Kruskal-Wallis H非参数检验比较各组主观图像质量评分。结果 测得77、66和55和44 mAs组的辐射剂量分别为常规剂量组的70%、55%、40%和25%。感兴趣区1中77、66和44 mAs组与常规剂量组的SNR和CNR比较，差异均有统计学意义（F=24.018，20.386，P〈0.05）；77和66 mAs组的SNR和CNR高于常规剂量组，44 mAs组低于常规剂量组；55 mAs组与常规剂量组比较差异无统计学意义（P〉0.05）。感兴趣区2中77、66和55 mAs组SNR和CNR高于常规剂量组，4组差异均有统计学意义（F=15.934、13.818，P〈0.05）；44 mAs组与常规剂量组比较，SNR和CNR差异无统计学意义（P〉0.05）。主观评价总体图像质量、图像对比度、血管边缘锐利度、主观噪声的结果，77、66和55 mAs组和常规剂量组得分较好，均≥ 3分，44 mAs组图像质量下降出现2分评价。5组图像的对比度、锐利度、主观噪声和总体图像质量差异均有统计学意义（H=10.268、14.542、15.840和11.014，P〈0.05）。受试者工作特征曲线（ROC）曲线分析显示55 mAs组为能够维持图像质量等同于常规剂量组的最适合的低剂量组，曲线下面积为0.844，95%置信区间为0.705～0.982。55 mAs组有效剂量为（3.545±1.396）mSv，约为常规剂量组剂量（9.271±2.892）mSv的40%。结论 迭代重建技术可以显著地降低噪声，在剂量降低为常规剂量FBP 40%的条件下仍能保持与之相当的图像质量。由此，可使患者上肢血管内瘘CT造影成像检查的辐射剂量降低至3.545 mSv。

Objective To assess the image quality（IQ） of an iterative reconstruction（IR） technique（sinogram affirmed iterative reconstruction, SAFIRE） for upper limb intravascular fistula CT angiography on a second generation dual-source CT scanner and determine the optimal dose level using IR with IQ comparable to filtered back projection（FBP）. Methods The experiment was a prospective study. A total of 150 patients were averagely divided into 5 groups to perform upper limb intravascular fistula CT angiography. The control group was scanned with conventional tube output （120 kV, 110 mAs） and images were reconstructed with FBP. The tube current of the other 4 groups（reconstructed with SAFIRE） was successively reduced to 70%, 60%, 50% and 40% of that of control group. Quantitative measurements of signal-to-noise ratio （SNR） and contrast-to-noise ratio （CNR） were obtained in two regions of interest （ROI） in each group. The ROIs were placed at the subclavian artery near shoulder joint, and the stoma of radial artery and cephalic vein. Analysis of variance （ANOVA） was used for comparisons of objective evaluation parameters（SNR, CNR） and radiation dose （CTDIvol, DLP, E） among 5 groups. Two radiologists graded IQ in a blinded fashion on a 4-point scale （4-excellent, 3-good, 2-fair and 1-poor）. The Kruskal-Wallis nonparametric test was performed for detection of differences in subjective evaluation of IQ among groups. Results The radiation doses of groups 77, 66, 55 and 44 mAs were 70%, 55%, 40% and 25% of the control group respectively. The differences of SNR and CNR among groups 77, 66,44 mAs and the control group were statistically significant in ROI1（F=24.018,20.386,P〈0.05）, SNR and CNR of group 77 and 66 mAs were higher than the control group while group 44 mAs was lower. The difference of SNR and CNR among group 55 mAs and the control group was not statistically significant in ROI1（P〉0.05）. The differences of SNR and CNR among groups 77, 66, 55 mAs and the control group were statistically significant in ROI2（F=15.934,13.818,P〈0.05）, and groups 77,66 and 55 mAs had higher SNR and CNR. The difference of SNR and CNR between group 44 mAs and the control group was not statistically significant in ROI2（P〉0.05）. Groups 77, 66, 55 mAs and the control group had higher scores of the subjective IQ（≥ 3） which was descended in group 44 mAs. The differences of contrast, sharpness, noise and general impression of images among the 5 groups were statistically significant by Kruskal-Wallis nonparametric test（H=10.268,14.542,15.840,11.014,P〈0.05）. Using the criterion of excellent IQ（score 4）, the ROC curve of dose levels and IQ acceptability established a reduction of 60% of effective dose[Group 55 mAs：（3.545±1.396）mSv] as an optimum cutoff point（AUC：0.844,95%CI：0.705-0.982）. Conclusions Iterative reconstruction technique could provide a dose reduction to 40%（3.545 mSv） compared with the filtered back projection in upper limb intravascular fistula CT angiography without reducing equivalent image quality.